You are on page 1of 5

HR ANALYTICS

Assignment 1 – Solve HR Problem using Machine Learning on R

Shamik Chakraborty

MBA(BA)
21020343061
Problem Statement and Model Data Acquisition, Understanding Model Fitment and Lessons
Introduction
Selection and Visualization Learnt

Introduction to HR Analytics Major Trends in HR Analytics

• Data driven process used by Human Resources Department to optimize • Some organisations have become good at people analytics but most are
human capital (workforce) management not there yet.
• Also known as People Analytics, Workforce Analytics and Talent Analytics
• By collecting and analyzing critical HR data, companies can generate • HR Analytics is better at delivering insights to HR leaders than to other
actionable insights and improve workforce, people and talent management stakeholders
performance.
• Organisations mostly rely on descriptive HR analytics rather than on the
Importance of HR Analytics more sophisticated predictive and prescriptive analytics

 Can be leveraged to create an iterative framework for making necessary • The most difficult elements of people analytics are data integration,
refinements in order to remain competitive and meet new challenges in cleaning and visualization
today’s business world
 Can address specific challenges such as: • Compensation is most commonly cited as the functional area where HR
• why there is a sudden spike in employees who quit after 2 years work Analytics are most important
anniversary
• why the quality of applications have fallen in the recent months. • HR professionals are most likely to say that an interactive visualisation is
the best way to present HR analytics

Fig 1. The HR Analytics


Management Cycle
Problem Statement and Model Data Acquisition, Understanding Model Fitment and Lessons
Introduction
Selection and Visualization Learnt

Problem Scope Problem Statement

• Challenges is Salary Negotiation as part of talent acquisition: • Prediction of optimum salary level based on experience and level
• Spike in salary expactations from the prospective candidates
in the job change • Prediction using ‘Fit-for-purpose’ regression model

• Inaccurate and fake salary information shared by the • Visualization of the model to check for the model fit with respect
candidates, which are incommensurate with their relevant to the input training data
experience
• Calculation of the model score for future reference to take
• Building an intelligent solution to quickly reconcile and decisions
validate the Salary Ask vs the Experience Level

Model Selection
• The predictor variables are Entry_Experience and Level.
• Multiple Linear Regression
• Regression helps to quantify the relationship between one or
Why Regression Model? more independent variables and a response variable

• Regression analysis is a statistical technique to determine the • Thus we fit a regression model to predict the salary of future
relationship between a single dependent variable and one or observations
more independent (predictor) variables.
• This can help the model for future data to predict the optimum
• The analysis yields a predicted value for the criterion resulting salary.
from a linear combination of the predictors.
Problem Statement and Model Data Acquisition,Understanding Model Fitment and Lessons
Introduction
Selection & Visualizing Learnt

Data Acquisition Visualising the Data


• Data acquisition meaning is to collect data from relevant sources before it • Data visualization is the graphical representation of information and data
can be stored, cleaned, preprocessed and further mechanisms.
by using visual elementsto understand trends, outliers and patterns in
data
• It is the process of retrieving relevant business information, transforming
data into required business form and loading it into designated systems.

• In this case,

Imported dataset  IT_Services_Salaries.csv


Checked for null values
Dropped column ‘Positions’ as part of transforming data

• Such data can be acquired from secondary and public databases pertaining
to particular company through sources like glassdoor, ambition box etc

Understanding the Data

• The goal is to gather insights from the data which would be helpful for
further analysis

• The dataset comprises of 4 columns:


 Position Fig 2. Scatter plot to demonstrate linear
 Entry Experience relationship between Entry_Experience and
 Level Annual_Salary
 Annual Salary
Problem Statement and Data Acquisition, Iterative Model Fitment and
Introduction
Model Selection Understanding and Visualizing Lessons Learnt

Iterative Model Fitment

• Model fitting is a calculation of how well a machine learning model generalizes related
data to that from which it has been taught.

• Iterative modelling allows machine learning models to improve accuracy

Predicting New Data

• The model has R square of 0.9810 and Adjusted R-square of 0.9735 i.e the model can
predict very closely to the actual expected output.

R square: 0.9810 Lessons Learnt


Adjusted R-squared: 0.9735
• Predicting salary amount based on certain features

• Linear regression to make predictions for continuous/numerical variables.


Fig 3. Plot of Actual vs Predicted Salary Values

You might also like